Bridge2AI at Year 4: The Operational Reality of Ethical, AI-Ready Data Generation

Presentation Type: Panel

Click to View Presentation

Presentation Time: 10:15 AM - 11:30 AM

Primary Track: Data Science/Artificial Intelligence


While a previous AMIA session introduced the vision of the NIH Bridge2AI consortium, Year 4 marks the transition from architectural design to tangible impact. This panel presents the consortium’s pioneering flagship data releases, demonstrating that AI-readiness is not merely a technical standard but a sociotechnical achievement. Panelists representing the four Bridge2AI Grand Challenges will showcase >30TB of newly available multimodal data, highlighting how Ethical Scaffolding and Team Science were operationalized to address critical Implementation bottlenecks. We will examine specific utilization cases: how privacy ethics drove the development of watermarking tools for diabetes data; how transdisciplinary teaming enabled the harmonization of 1.6 billion EHR rows across 14 hospital sites; how the SAFE ethical framework shaped the release of voice biomarkers; and how robust data and tools standards help address “black box” AI/ML concerns. This session offers a practical roadmap for researchers to access, utilize, and trust these ethically sourced resources.

Moderator:
Mónica Muñoz Torres, PhD
University of Colorado Anschutz

Speaker(s):
Aaron Lee, MD, MSCI
Washington University in St. Louis.

Satrajit Ghosh, PhD
Massachusetts Institute of Technology

Mónica Muñoz Torres, PhD
University of Colorado Anschutz

Yulia Levites Strekalova, PhD
University of Florda

Timothy Clark, Ph.D.
University of Virginia

Author(s):
Satrajit Ghosh, PhD - Massachusetts Institute of Technology; Aaron Lee, MD, MSCI - Washington University in St. Louis; Yulia Levites Strekalova, PhD - University of Florda; Timothy Clark, PhD - University of Virginia; Mónica Muñoz Torres, PhD - University of Colorado Anschutz;

---